[Video] Office Hours Off the Coast of Norway


Come join me in my cabin aboard the Hurtigruten Nordlys, sailing up the coast of Norway, and we’ll cover your top-voted questions from https://pollgab.com/room/brento.

Here’s what we covered today:

  • 00:00 Start
  • 00:57 Accidental DBA: What are the best options you have seen for database deployment rollbacks with releases (if issues were encountered)? My understanding is that restoring to a pre-deployment backup or snapshot would break the transaction log chain so wondering if there’s better options.
  • 02:30 Hera H: Which VM technologies should the SQL DBA know about for optimizing SQL VM performance in Azure?
  • 02:51 END TRY BEGIN CRY: Hi Brent. For large text columns, like Notes, that are not going to be indexed, is it better to have a VARCHAR(MAX) so that the data is stored “off page” or is it still better to have it as a VARCHAR(2000) for example, or does it make no difference. Thanks
  • 04:37 Ryan: We admin a CRM server and from time to time we see high cpu and a high wait on resource semaphore. Where should we start? We feel like we have tried everything.
  • 05:30 Thomas Franz: My msdb on SQL 2022 Enterprise is in FULL recovery mode (because of job_history etc.). When I install a new CU it will always be set to SIMPLE recovery. Any idea why? Are there known problems with msdb and FULL recovery or may this be a Microsoft bug?
  • 07:45 Tony F: Did many people wear masks on the plane? Did you feel safe?
  • 09:34 Robert McTables: You wrote a great post a few years ago (How Much Memory is “Normal” for SQL Servers?), do you have any new data trends that you could share about newer SQL versions? Did the trend continue of starving older versions?
  • 10:07 Wicket: Hi Brent, why does SQL Server use the clustered index key in non-clustered indexes instead of the RID? A clustered key could be many wide columns, presumably there are benefits this way but why not just always use a RID and mitigate large keys bloating indexes?
  • 10:54 Red Utley: What is your opinion of using Windows S2D for hosting Azure SQL VM storage?
  • 11:26 Ozan: Hi Brent, from security perspective, would you always enable TDE, since otherwise data can be read in plain text via hex editor as you already demonstrated? Or is there another way to encrypt the data(file)? Thanks
  • 14:04 About Hurtigruten cruises
Previous Post
Something’s Up with Parallelism in Azure SQL DB Serverless
Next Post
[Video] Office Hours at the Arctic Circle

2 Comments. Leave new

  • >aboard the Hurtigruten

  • Michael Roland Lemay
    October 5, 2023 3:39 pm

    This is for Thomas Franz: It sounds like the reason you have MSDB in Full recovery is to preserver data that’s maintained there. In lieu of that you could just create another database on that server and run a job every night to grab the data you need out of MSDB and maintain your history there instead.


Leave a Reply

Your email address will not be published. Required fields are marked *

Fill out this field
Fill out this field
Please enter a valid email address.